Finding an effective and efficient method for teaching speech sounds has been of the great interest in language related fields. There are many consonant training studies, but there is limited research on vowel training. Since vowels convey not only segmental but also suprasegmental information, it is necessary to improve both perception and production of vowels in order to increase one's communicative efficiency. The present research is on training methods for vowels acquired in a second language. Trainees will be learners of American English (AE) who have a common native language. The first purpose of the present research is to evaluate the impact that different ranges of stimulus sets used in a perceptual training have on the effectiveness of training. A pre test post test design will be employed to gauge change in trainees' performance resulting from training. The second purpose of this study is to assess the transfer of perceptual training to production. All trainees will record their speech before and after training. Recorded speech samples will be acoustically analyzed as well as evaluated by native speakers of AE for intelligibility and accentedness. The third purpose of this study is to assess the possibility of long-term retention of the training effects. These results will provide empirical support for a theoretical framework for the speech perception-production link. These results can also be used to extend the research to propose new vowel training methods.